Jump to content


Photo

Guiding or flexure issue?

  • Please log in to reply
72 replies to this topic

#51 Jon Rista

Jon Rista

    Vanguard

  • -----
  • Posts: 2281
  • Joined: 10 Jan 2014
  • Loc: Colorado

Posted 26 March 2014 - 08:32 PM

Well, if you are already using scope rings to support the focuser/drawtube then that should help. But, if the rings are just supporting the optical tube then you still need to be concerned about flexure in the focuser/drawtube itself (the loose slip fit between the tube and the focuser housing could be causing problems). I don't know which setup you are using. Rings can hold the optical tube fairly steady, but that alone won't prevent the drawtube from moving.


I'll take some pictures of the ST80 setup. The rings currently just hold the optical tube. I hadn't thought of using the rings to hold the drawtube...I guess I could get a third smaller ring to do that. The focuser has a locking knob that presses on the drawtube. When I tighten it down, it does change the framing a bit, but it eliminates most of the play in the drawtube. It doesn't eliminate all of it, and I understand that all it takes is a fraction of a millimeter to cause problems.

As for the Orion mini-guider, it's not perfect but my major concern with it would be the image scale it produces with your Orion Starshoot Guider. I mentioned early on that it was probably nearing the boundary that I would consider safe for an autoguider image scale (1:5 with your setup), but that it was probably not the major source of your problems. To repeat from earlier, Craig Stark says that the star centroid calculations in PHD can determine a star's center to an accuracy of 1/5 pixel under even very noisy conditions (or much better under less noisy conditions, better seeing, better guider, etc.).

However, there are some in the industry who doubt that the mathematical simulations that are being used to PREDICT such centroid accuracy are really that good in the actual field. I guess one "rule" that has been kicked around is a 1:30 scale, but given the above I find that perhaps very optimistic. So, YMMV when moving much beyond that 1/5 pixel accuracy (IMO).


Aye! This is what you mentioned on my previous thread on this topic. My ratio with the 50mm guidescope is 1:4.5. So, while it isn't beyond that 1:5 ratio that Stark mentions, it's pretty close. At best, I was able to get about 4-5" P2P tracking with the 50mm, with RMS on average around 2", but as high as 3". At the absolute best, when pointed closer to the meridian, with everything tuned about as optimally as I currently have the skill to tune it, the 50mm guidescope and SSAG got as low as 1.5" RMS, however my P2P error was still about 4", so while guiding seemed a bit smoother, the effect on my star size and quality did not really change much.

With the ST80, the ratio is 1:1.8. If I can resolve my flexure issues, I think I can take my guiding performance much farther with the ST80 than I can with the 50mm mini guidescope. It was this 1:5 ratio thing, along with the mention in the previous thread that the 50mm mini guidescope had flexure issues of its own, that made me buy the ST80.

I went back and looked at your most recent unguided PHD graphs and I see two things. First, it appears that you have an error in one axis of your polar alignment. I say this because your drift lines are much better on the meridian than they are toward the west. Specifically take notice of this statement from the following document/link:

http://celestialwond...gnmentError.pdf

Notice that the optimal point for observing azimuth error is at the intersection of the equator and the meridian. But due to the orthogonal geometry this is also the location where altitude error is virtually undetectable. Conversely, we observe altitude errors best on the equator at the eastern or western horizon, and in this location azimuth error is minimized. This is how we are able to use the drift method to measure and correct azimuth and altitude errors independently. Of course, this all assumes that the mount base is leveled, as an unleveled base would cause azimuth corrections to affect altitude and vice versa.


Interesting. Just to make sure you are fully aware (I'm not sure if this changes anything or not), but my unguided PHD altitude graph was not exactly "near the horizon". I have homes and trees in the way of that. I think I was probably about 45° up from the horizon. I do agree, there is more error in altitude than azimuth. I'm a bit wary of tweaking altitude, because there is no smooth way to do it. The weight of everything on the mount, which ultimately sits directly on that one L-bolt for altitude adjustment, makes it difficult to make fine adjustments. Maybe it's just a skill that you learn with practice, but usually when I try to adjust altitude, I end up making short jumps in movement, which results in the alignment swinging even farther out, rather than tighter in, and the night usually devolves into altitude L-bolt tweaking instead of imaging.

PHD shows my altitude error to be pretty low, though...0.29', or 17.4". Is that enough to be concerned about?

Secondly, there seems to be a periodic shift in declination that follows the PE in R.A (on the unguided graph taken near to the meridian). I see this same issue on my Celestron AVX mount and I THINK this is being caused by a wobble in the polar axis that happens as the R.A. worm gear moves against the polar spur gear. This could happen if there is any play in the bearing on the polar axis and if the R.A. worm and spur gear are not in perfect alignment (among other possibilities). It could also be affected (made worse or better) depending upon the balance of the mount.


Aye! This is the declination oscillation issue. I really don't know what causes it, any insight into what might is extremely welcome. It's been another thorn in my side, and has made it practically impossible to guide in dec because dec drift will periodically swing to the non-guided side of the midline in PHD, and then he potential error is uncontrolled...it often gets as high as 3-4" off the midline before finally drifting back down. I'd tried guiding dec in both directions, and of course backlash then becomes the problem, and longer and longer guide pulses eventually send declination racing to the opposite side, and you get a faster and more pronounced oscillation.

It's an infuriating problem, however it doesn't affect my stars as much as the flexure issue does, so I've just lived with it for now. And, rather ironically, when my declination plot is smooth, the elongation of the stars in RA tends to be much more stark and less easy to work around...so as long as the unguided swings don't top 2", I've usually welcomes it. :p

I'd be interested if anyone else has experienced such an issue and what they think is causing the APPARENT correspondence between the PE and a similar period/cycle of movements on the declination axis (which when unguided should not be moving).


I too would be interested in hearing anyone else's experience with this.

#52 james7ca

james7ca

    Viking 1

  • -----
  • Posts: 542
  • Joined: 21 May 2011
  • Loc: San Diego, CA

Posted 26 March 2014 - 09:02 PM

PHD shows my altitude error to be pretty low, though...0.29', or 17.4". Is that enough to be concerned about?

That wouldn't be any problem at all if it is an accurate measure of your polar alignment.

I think it is unlikely, however, that that is the true error. I don't know how accurately PHD2 reports polar alignment error, but there are other ways to check the quality of your polar alignment. You might also want to open your PHD log in something like PHDLab and see if the two programs agree on the error, you just need to select a DRIFT entry, not one that has had guiding enabled.

PHDLog is a freeware program, available on the web.

#53 Jon Rista

Jon Rista

    Vanguard

  • -----
  • Posts: 2281
  • Joined: 10 Jan 2014
  • Loc: Colorado

Posted 26 March 2014 - 09:20 PM

Here are some photos of the ST80 setup. I've annotated a few of them. I actually messed with the drawtube locking nut, and when it's fully tightened down, there is no play in the drawtube at all. I've applied quite a bit of force on it. It's long, especially with the t-thread extension, so there still just be flex in the materials themselves, but I do not believe that where the drawtube moves into the focuser is going to be a significant source of flex itself.

Posted Image

Posted Image

Posted Image

Posted Image

I have PHDLab, which is also capable of showing the drift rate. I'll run my logs through that and see what it says.

#54 Madratter

Madratter

    Fly Me to the Moon

  • *****
  • Posts: 6212
  • Joined: 14 Jan 2013

Posted 26 March 2014 - 10:19 PM

Another way of measuring polar alignment error is using AstroTortilla. I was working on refining my polar alignment with it tonight. But polar alignment can be embarrassingly far off and as long as the mount can guide out the problems, it doesn't make that big a difference except in quite long exposures. A couple minutes is not enough to make a difference.

Attached Files



#55 Jon Rista

Jon Rista

    Vanguard

  • -----
  • Posts: 2281
  • Joined: 10 Jan 2014
  • Loc: Colorado

Posted 26 March 2014 - 10:35 PM

Yeah, I've used AT to check my PA before. I should do that next time I have a clear sky.

In the mean time, here is what PHDLab says:

Altitude:
Posted Image
(Full size: http://i.imgur.com/1ojOAVF.jpg)

Azimuth:
Posted Image
(Full size: http://i.imgur.com/fQt5NtR.jpg)


I did notice something interesting when I first loaded my guide logs into PHDLab. It showed in the Guider specs tab that my guiding focal length was 777mm, instead of 400mm. At 777mm, the numbers PHDLab spit out for DEC drift and Polar Align Error were closer to what PHD said they were. However, once I set my guider and imager specs properly, the numbers changed significantly. If PHDLab is correct, my Altitude PAE is over 3', which is many times worse than I thought it was.

I guess that would explain the rouge tracking in RA when I point west.

#56 james7ca

james7ca

    Viking 1

  • -----
  • Posts: 542
  • Joined: 21 May 2011
  • Loc: San Diego, CA

Posted 26 March 2014 - 10:39 PM

I'm not going to outright suggest that you need rings around the focuser drawtube, but I've seen setups where some have done just that. However, I would not try that unless you have good reason to think that could be a problem (or as a near final and relatively cheap modification if everything else fails). One reason I could see where you wouldn't want to put rings there is because it could cause stress flexure under temperature change (from different rates of expansion or contraction between the tube, rings, and dovetail). Of course, that could happen anywhere or even without additional rings but it could be yet another change that either does nothing or potentially could even make things worse.

I guess it would also depend upon how quickly the temperature is changing. If everything is temperature equalized and the temperature doesn't change during the course of your exposure then there shouldn't be any expansion or contraction in the parts. However, over longer periods of time as the temperature changes you will most likely see some shifts caused by expansions or contractions (as well as focus changes, which can be another problem).

Once again, I'd suggest that you take a look at the link I gave you earlier about finder/guider flexure. It shows how to test it and gives graphic examples of how bad (or good) it can be.

#57 james7ca

james7ca

    Viking 1

  • -----
  • Posts: 542
  • Joined: 21 May 2011
  • Loc: San Diego, CA

Posted 26 March 2014 - 11:01 PM

With PHDLab, I'm pretty sure that you need to select "Drift check" under the popup menu rather than "Guiding" if you are looking at a record that was done with the guiding outputs disabled. This will give you a better estimate on your polar alignment error. You also want to input the declination of the guide star so that you can see how much field rotation you'd have with the given setup (you change this under the "Positioning" tab).

In any case, these are only ESTIMATES of what your polar alignment error might be. In fact, even the guided PHD logs are only an estimate of what MAY be happening with your imaging. The only way to know for sure is to look at the actual imaging results and it may be worth running some tests where you use very short exposures and then look at the registration results from the images. I don't know if DSS will output the registration errors, but some software (like PixInsight) will log the registration errors so you can see exactly what is happening at the image plane. PixInsight also allows you to measure FWHM and eccentricity which will give you a much more quantitative measure on just how good the images are (rather than just trying to eye-ball the results and express an opinion).

#58 Jon Rista

Jon Rista

    Vanguard

  • -----
  • Posts: 2281
  • Joined: 10 Jan 2014
  • Loc: Colorado

Posted 26 March 2014 - 11:25 PM

Ah, after switching to Drift Check, and setting declination for both unguided sequences, I am now getting the following:

Altitude:
DEC Drift: 0.23"/min
PAE: 0.89'

Azimuth:
DEC Drift: 0.06"/min
PAE: 0.24'

That is a little more than PHD was saying, but not enough of a difference to be an issue, I think.

Regarding the flexure paper...is that the Borg mini finder paper you linked?

I've requested a trial license for PI, I guess (as seems to be the MO among astro software developers) that they manually process license requests, so I am not sure when I'll actually get the license key. I'm looking forward to using it, though...sounds like an incredibly powerful tool.

#59 james7ca

james7ca

    Viking 1

  • -----
  • Posts: 542
  • Joined: 21 May 2011
  • Loc: San Diego, CA

Posted 26 March 2014 - 11:32 PM

Regarding the flexure paper...is that the Borg mini finder paper you linked?

Yes.

#60 Jon Rista

Jon Rista

    Vanguard

  • -----
  • Posts: 2281
  • Joined: 10 Jan 2014
  • Loc: Colorado

Posted 26 March 2014 - 11:45 PM

Thanks. That flexure test is really easy. I may just haul my mount out to Cherry Creek (a nature reserve nearby, where there is plenty of open space and a good view of distant objects), and perform that same test and see what the results are (for both my guider and my imager.) For the rate at which star trails grow in my imaging, I am pretty sure I probably have at least 3-4 times as much drift from flexure as the author of that article (seems he had ~800mm scope @ 2"/px, where as I have a 600mm scope @ 1.48"/px, and my drift is much more pronounced than his).

#61 proteus5

proteus5

    Apollo

  • *****
  • Posts: 1360
  • Joined: 24 Jun 2010
  • Loc: Southern New Jersey

Posted 27 March 2014 - 05:39 AM

I'm certainly in no way knowledgeable about flexture, but reading through this is very interesting. One thing that popped into my head was as a test, would it be possible to mount your st-80 directly to the Atlas via a dovetail, and image through it, while using the SSAG with the 50mm guide scope mounted in the finder bracket of the st-80 for guiding? This would eliminate your lens, and the ADM dual mounting bar from the setup. I'm sure this might induce any number of other guiding issues since as you have found the focuser of the st-80 isn't the best, but if the issues are even a little similar, that might point to the mount as the problem, and if they go away it would at least rule out the mount. Just a thought, as it seems a you have all the parts for such a test.

#62 JJK

JJK

    Surveyor 1

  • -----
  • Posts: 1950
  • Joined: 27 Apr 2008

Posted 27 March 2014 - 08:37 AM

Drift due to differential flexure needn't be random. For example, if the cables attached to your guidescope tug on it, you'll get a more deterministic movement.

#63 JJK

JJK

    Surveyor 1

  • -----
  • Posts: 1950
  • Joined: 27 Apr 2008

Posted 27 March 2014 - 08:42 AM

Just a question, why not get a get a nice triplet in the focal length that was made to be used for this? Then no worry about an expensive lens if you need to secure it better, add rings, etc. (I've beat the hell out of 300 and 400 2.8's and 600 4's but they always belonged to the publication so it was expected, LOL. But if it were my own...)


Or an Officina Stellare Veloce RH200.

#64 Thirteen

Thirteen

    Messenger

  • -----
  • Posts: 433
  • Joined: 12 Jul 2013
  • Loc: Milford, Michigan

Posted 27 March 2014 - 08:54 AM

I'm certainly in no way knowledgeable about flexture, but reading through this is very interesting. One thing that popped into my head was as a test, would it be possible to mount your st-80 directly to the Atlas via a dovetail, and image through it, while using the SSAG with the 50mm guide scope mounted in the finder bracket of the st-80 for guiding? This would eliminate your lens, and the ADM dual mounting bar from the setup. I'm sure this might induce any number of other guiding issues since as you have found the focuser of the st-80 isn't the best, but if the issues are even a little similar, that might point to the mount as the problem, and if they go away it would at least rule out the mount. Just a thought, as it seems a you have all the parts for such a test.


Agree. As long as it minimizes the effect, you could then start to rebuild the setup piece by piece and find whatever causes the tailing.

#65 proteus5

proteus5

    Apollo

  • *****
  • Posts: 1360
  • Joined: 24 Jun 2010
  • Loc: Southern New Jersey

Posted 27 March 2014 - 09:11 AM

Just another odd thought. I wasn't able to see the pictures last night when I posted as I was at work and they were blocked. Looking at this picture, would it be possible that the thumb screws labeled "these do nothing" could be slightly intruding into the light path causing some slight diffraction which is confusing PHD? If so could that explain the regularity of the error. I don't know if this theory is even possible but thought I'd throw it out there.

Attached Files



#66 Jon Rista

Jon Rista

    Vanguard

  • -----
  • Posts: 2281
  • Joined: 10 Jan 2014
  • Loc: Colorado

Posted 27 March 2014 - 12:41 PM

@Robert: Good thought. I had it too. ;) If you look closely, you will see that those thumb screws are actually unscrewed about a far as they can go, without actually being removed. I did not want to introduce a source of stray light, but I also did not want them to enter into the light path.

#67 Jon Rista

Jon Rista

    Vanguard

  • -----
  • Posts: 2281
  • Joined: 10 Jan 2014
  • Loc: Colorado

Posted 27 March 2014 - 12:44 PM

Drift due to differential flexure needn't be random. For example, if the cables attached to your guidescope tug on it, you'll get a more deterministic movement.


Back when I first started, it's likely that cables were part of my problem. These days, I have my cables fairly well under control. I've tied them down to the ADM DSBS, from where they hang down and drape close to the mount itself. Under normal tracking, there is very little movement on the cables in general, and where they plug into the cameras, there shouldn't be any movement at all (not, at least, from the movement of the mount). They aren't pulled tight, but neither are they loose.

#68 james7ca

james7ca

    Viking 1

  • -----
  • Posts: 542
  • Joined: 21 May 2011
  • Loc: San Diego, CA

Posted 27 March 2014 - 09:10 PM

As for the flexure test (i.e. Craig Stark's procedure), another person mentioned earlier that you could do that under a night sky. Just align both scopes to a star and then swing the scope to the other side of the meridian and check to see if both are still optically aligned. You should also be able to use plate solving software to find the centers of both scopes/lenses and then move across the meridian and check them again (for their plate-solved centers).

#69 Jon Rista

Jon Rista

    Vanguard

  • -----
  • Posts: 2281
  • Joined: 10 Jan 2014
  • Loc: Colorado

Posted 27 March 2014 - 09:58 PM

@James: I have a quick question, regarding that test. I have some cone error due to the way my lens attaches to the mount. The screws that fit the dovetail had to be the smaller thread size, so I had to use the thread adapter on the lens foot. That left about a 3/4mm lift for that particular screw where it attaches to the dovetail. I've tried to shim with some layers of aluminum foil, and I did manage to correct some of my cone error...but I am actually unable to do a meridian flip and even get the same star in my FoV. It's close (just outside the frame, I think), but that itself would be the single largest source of error in doing this particular test.

I've used AstroTortilla to plate solve, however it only solves well enough for one side of the meridian...it doesn't seem to deal with cone error, so every time I flip, I have to plate solve again. I'm also not even sure my pointing accuracy is good enough. I once tried to polar align with AstroTortilla...due to the inherent pointing error of the mount, I ended up throwing my PA way out of alignment, because I can't point back and forth between two stars and have them end up in the exact same place within the FoV each time...the error accumulates, and eventually your just way off, and totally misaligned.

Is there any way I could perform this test without having to first figure out how to ideally correct my cone error? I'm not even sure if the lens foot itself is even level, such that if I perfectly dealt with that 3/4mm error introduced by the thread adapter whether I would still have some cone error or not. It could take endless hours of shimming and testing and shimming and testing and unshimming to correct my cone error.

#70 Madratter

Madratter

    Fly Me to the Moon

  • *****
  • Posts: 6212
  • Joined: 14 Jan 2013

Posted 27 March 2014 - 10:04 PM

Jon, all you need to do is take one set of shots with the counterweights East high, another with them west high, and for good measure ones where the counterweights are down. You take the shots with both the guide camera and the main camera.

Then plate solve all the shots.

Then figure the distance between the centers of the pairs of shots.

You don't actually have to be pointed at the same target, although you can.

#71 josh smith

josh smith

    Viking 1

  • *****
  • Posts: 905
  • Joined: 13 Oct 2013
  • Loc: Pittsburgh, Pennsylvania

Posted 27 March 2014 - 10:22 PM

All you need is the difference in the differences. Hence; the differential flexure. :) if you do this without moving in dec at all, you can determine your flexure/drift rate as well and figure out how much of the trailing in the stars is due to differential flexure and how much is due to guiding/other issues.

#72 Jon Rista

Jon Rista

    Vanguard

  • -----
  • Posts: 2281
  • Joined: 10 Jan 2014
  • Loc: Colorado

Posted 27 March 2014 - 10:32 PM

So, I only need to take images to figure out differential flex, then compare the differences of those. There is no value in determining the total flex of the system as a whole? (I.e. the difference in flex between peirside east counterweight level, to pierside west counterweight level, with declination at some fixed position (say 0°)?)

#73 Madratter

Madratter

    Fly Me to the Moon

  • *****
  • Posts: 6212
  • Joined: 14 Jan 2013

Posted 27 March 2014 - 10:35 PM

There is some value in figuring out the actual figure. But actually, you could do that just using the non dithered sequence you already have.






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics